Transactions of JWRI, Vol.40 (2011), No.

Effect of Heat Cycle in Multi-layer Welding on Charpy Absorbed

Energy in Low and High Toughness Steel

SAKINO Yoshihiro* and KIM You-Chul**

Abstract
In this study, the heat cycle is determined in the case of multi-layer welding using three-dimensional
heat-conduction analysis. The relationship between the heat cycle of multi-layer welding and the absorbed
energy is defined by simulating various heat cycles of multi-layer welding by performing the synthetic HAZ
test using steel with a relatively high Charpy absorbed energy and steel with a relatively low Charpy absorbed
energy. On the basis of this relationship, the authors have studied the effects of heat cycle conditions on the
Charpy absorbed energy of weld interfaces. In the results, the Charpy absorbed energy increased more than
in the case of single-pass welding in some heat cycle in the case of multi-layer welding, both in
high-toughness steel and low-toughness steel. The Charpy absorbed energy of the weld interface was greatly
susceptible to the achieved temperature after the maximum achieved temperature, and the range of the
temperature after the maximum temperature was achieved in which the Charpy absorbed energy was found to
be different depending on the steel toughness.
KEY WORDS:(Welding heat affected zone) (Synthetic heat-affected zone test) (Charpy absorbed energy)
(Multi-layer welding) (Heat-conduction analysis)

of factors such as the displacement of the notch tip from

the desired position and the progress of cracks outside
the area of analysis. As described above, it is difficult
to determine the local toughness of HAZs using actual
welding materials. A potential solution could be to
adopt a synthetic HAZ test for estimating toughness in
which HAZs are simulated using a simulated
thermal-cycle device. The simulated thermal cycle
device can simulate various heat processes using rapid
high-temperature heating by applying an electrical
current, and it can simulate the welded material after
recrystallization as homogeneous materials. When a
homogeneous material is used, variations resulting from
heterogeneity of material properties can be eliminated,
and local portions to be analyzed can be tested precisely
using various evaluation tests such as the CTOD test and
the Charpy impact test. The synthetic HAZ test was
also used to test the strength properties, which are as
important as the Charpy impact properties in
characterizing HAZs, to examine the relationship
between the strain required for initiating ductile cracks
and the material properties1), and to determine the tensile
strength at high-speed loading2).
The authors focused on 400-MPa-grade and
490-MPa-grade structural steels as test materials, and

1. Introduction
Extensive study of the required toughness of base
materials and weld metals has been conducted through
full-scale and model experiments in order to prevent
brittle fractures of beam-to-column connections and
improve their deformation capacities. Heat-affected
zones (HAZs) formed as a result of welding where
toughness may be very low are also attracting attention,
and research has been conducted on specimens that have
been cut out from actual HAZs. At present, the Charpy
absorbed energy is widely adopted as a factor that can be
controlled to prevent brittle fractures. The Charpy
absorbed energy at 0 C of 27 J and 47 J, which are
required for steel and weld metals in order to prevent
low-stress brittle fractures, and 70 J, (which is proposed
as the value required for preventing general yield brittle
fractures), have been used in Japan. However, unlike
base materials and weld metals, in HAZs, large changes
in the material properties are observed in a small area.
Therefore, the toughness test methods, such as the crack
tip opening displacement (CTOD) test and the Charpy
impact test often fail to produce accurate results because
Received on December 26, 2011
* Assistant Professor
** Professor

Transactions of JWRI is published by Joining and Welding

Research Institute, Osaka University, Ibaraki, Osaka 567-0047,
Japan

69

Effect of Heat Cycle in Multi-layer Welding on Charpy Absorbed Energy

Table 1 Tensile test results and chemical composition

Tensile test results
YR
Y
U
H
(%)
(MPa) (MPa) (%)
High-toughness
steel
Low-toughness
steel

Si Mn
10-2

Chemical composition (mass%)

P
S Cu Ni Cr V Mo N
Ceq
10-3
10-2
10-3 10-2

279

427

34

65

13

21

91

13

30

281

456

31

62

24

39

20

32

Ceq=C+Si/24+Mn/6+Ni/40+Cr/5+Mo/4+V/14
multi-layer welding.

simulated CGHAZs, whose toughness is often the lowest

among all HAZs, as homogeneous material by the
synthetic HAZ test. These CGHAZs were used in the
Charpy impact test to determine the Charpy absorbed
energy.
This enabled the demonstration of the
relationship between the equivalent welding-heat input
obtained from the cooling time according to the
simulated heat cycle and the Charpy absorbed energy.
It also allowed their comparison with various standard
values, and facilitated a comprehension of the limit
values for welding-heat input when the CGHAZs exist as
a homogeneous material3). The authors also examined
the effects of pre-strain and aging on Charpy absorbed
energy at CGHAZs4). However, single-pass welding
has been the focus of this study for a considerable
period.
In this study, the heat cycle is determined in the
case of multi-layer welding using three-dimensional
heat-conduction analysis. The relationship between the
heat cycle of multi-layer welding and the absorbed
energy is defined by simulating various heat cycle of
multi-layer welding by performing the synthetic HAZ
test using steel with a relatively high Charpy absorbed
energy (hereinafter referred to as high-toughness steel)
and steel with a relatively low Charpy absorbed energy
(hereafter referred to as low-toughness steel)5). On the
basis of this relationship, the authors have studied the
effects of heat cycle conditions on the Charpy absorbed
energy of weld interfaces of two types of structural steels
with different levels of toughness, in the case of

In the experiment, SN400 (rolled steels for building

structure, 400-MPa-grade) flat steel (PL-22) was used as
the high-toughness steel, and SS400 (rolled steels for
general structure, 400-MPa-grade) H section steel
(RH-612 2021323) was used as the low-toughness
steel. Table 1 shows the tensile test results and
chemical composition analysis results of the sample
steels in the inspection certificate.
Figures 1 and 2 show the transition curves obtained
by performing the Charpy impact test on the sample
steels.
The absorbed energy at 0 C of the
high-toughness steel used in this study is about 185 J,
which is not a significantly large value. However, since
it is considerably larger than the values of 27 to 70 J,
standardized or proposed in JIS and in the literature
proposed by the Building Center of Japan, this steel is
called high-toughness steel in this study. The Charpy
absorbed energy at 0 C of the low-toughness steel is
very low (about 15 J), and its energy-transition and
fracture-transition temperatures are both very high
(approximately 70 C). CTOD tests using a notched
three-point bending specimens were also conducted, and
the results showed that the limit of the CTOD value at 0
C, c, was very low (0.0444 mm). As described so far,
it could be said that the toughness of the sample steel is
very low.

Crystallinity (%)

150

100

200

80

120

80

150

60

90

60

100

40

60

40

50

20

30

20

0
-80

-60

-40

-20

Absorbed Energy(J)

100

20
40
Temperature()

-40 -20

Fig. 1 Transition curve (high-toughness steel)

20

40

60

0
80 100 120 140 160 180 200
Temperature()

Fig. 2 Transition curve (low-toughness steel)

70

Crystallinity(%)

Absorped Energy (J)

Crystallinity (%)

250

Crystallinity(%)

Absorbed Energy(J)

Absorped Energy (J)

2. Outline of experiment

Transactions of JWRI, Vol.40 (2011), No. 2

electrodes. From the results of the hardness test and

from metallographic observation, it was confirmed that
the central area to which the heat cycle was applied was
a homogeneous material3). After the synthetic HAZ test
was performed, the specimens were processed to form
full-size V-notched specimens having dimensions of
101055 mm; these V-notched specimens were then
used in the Charpy impact test. A Charpy impact tester
with a capacity of 490 J was used, and the test
temperature was 0 C. Four pieces of specimen were
used for each condition.
When the heat cycle is known, the synthetic HAZ
test can simulate the HAZ that was affected by the heat
as a homogeneous material, but it is very difficult to
determine the heat cycle of an HAZ, especially when it
exists inside steel. In such cases, analytical approaches
are a possible option to determine the heat cycle.
Recent improvements in computing speed and memory
also enable heat-conduction analysis of multi-layer
welding of a butt weld zone as a full-scale
three-dimensional model and as a moving heat source
which the welding position continuously changes6).
Then, to determine the welding heat cycle at the
weld interface during multi-layer welding, a
three-dimensional unsteady-heat-conduction analysis of
the butt weld zone between the through diaphragm
(PL-23) and the beam (rolled H section steel RH-612
2021323 mm) flange was used. Figure 4 shows an
entire mesh drawing; Figure 5 shows the detailed
drawing of the weld zone in the A-A' section (cross

Thermocouple

Electrode

Electrode

Specimen

Fig. 3 Synthetic HAZ test

The simulated HAZ specimens of square-rod
-shaped (111160 mm) were cut out from the center of
the high-toughness steel and the low-toughness steel,
and a heat cycle was applied to the central area of these
specimens using a simulated thermal cycle device
(Gleeble 1500). The temperature was controlled using
a Type R thermocouple attached at the center of each
specimen, and helium was used as the atmosphere gas.
Figure 3 shows an image of the apparatus after the test.
The left-side electrode was moved away from the
specimen to allow it to be clearly seen in this image, but
was placed in the same manner as the right-side
electrode during the test to hold the specimen such that
there were 10 mm spaces between both sides of the
Diaphragm

Weld metal

Weld metal

Beam flange

Remarkable
element
Fig. 5 Detail of the weld zone
(A-A section)

Backing strip
Fig. 4 Analysis model at the butt weld zone between
the diaphragm and the beam flange

260

32

34

260

32

27

160

260

32

28

190

Pass
No.

Current Voltage Speed

(A)
(V)
(cm/min)

Fig. 6 Temperature dependence of the

physical constants7)

71

Interlayer
temperature
(qC)
100

Table 2 Welding conditions

Effect of Heat Cycle in Multi-layer Welding on Charpy Absorbed Energy

(Table 2). Figure 6 shows the temperature dependence

of the physical constants used in this analysis7). The
thermal efficiency is set at 0.8 because the welded area is
narrow.
In this study, the heat cycle at the location of an
element that is at the weld interface on the
first-layer-diaphragm-side of the A-A' section (the black
element indicated as Remarkable element in Fig. 5)
was determined. This section is the origin of the
first-layer welding, and it can thus cause weld defects
and may become the origin of brittle fracture.
Figure 7 shows the analysis results. The actual
welding was for 7 layers and 13 passes, but the figure
shows the results up to the third layer and the third pass
where the maximum achieved temperature is 500 C,
which is considered as lower limit temperature to
influence the structural changes of steels. This study
adopts the result shown in Fig. 7, which is obtained from
the model shown in Fig. 4, as an example of the heat
cycle of the beam-to-column connection, and examines
the effects of the heat cycle on the Charpy absorbed
energy of the HAZs of two types of steels with different
toughness during multi-layer welding.

Fig. 7 Results of three-dimensional

heat-conduction analysis
section of the region that is 10 mm inward from the
beam end where welding is started).
Only the
diaphragm (24024028 mm), the beam flange (200
25023 mm), and the backing strip (200249 mm) that
were in contact with the weld zone were modeled, and
the weld tabs were not considered. The welding is
represented as a moving heat source, and heat transfer to
the column and beam webs are not taken into
consideration.
The deposition sequence and the
welding conditions conform to the measurement record

3. Simulated welding heat cycle

and experiment results

1600

3.1 Comparison between single-pass welding

and multi-layer welding

1400

Temperature ()

1200

First, a synthetic HAZ test was conducted using the

heat cycle that modeled only the first pass (single-run)
and the heat cycle that modeled up to the third layer and
the third pass (multi-layer) based on the analysis results.
Figures 8(a) and (b) show the heat cycles used for the
test.
Figure 9 shows a comparison of the results of the
Charpy absorbed energy in the case of the single-pass
welding and those in the case of multi-layer welding.
This graph shows the values of the Charpy absorbed
energy of each specimen (high-toughness steel, H-D

1000
800
600
400
200
0
0

100

200

300

400

500

TIME (s)

(a) Single-run
1600

Absorbed energy (J)

1400

Temperature ()

1200
1000
800
600

250

H-D Steel
L-D Steel

200
150

400

100

200

70J
0
0

100

200

300

400

50

500

TIME (s)

(b) Multi-layer, High-Intermediate-Low

47J
27J
1

Virgin

Fig. 8 Simulated welding-heat cycle

(Comparison between single-pass welding
and multi-layer welding)

Single-run

Multi-layer

Fig. 9 Comparison of the Charpy absorbed energies

between single-pass welding and multi-layer welding

72

Transactions of JWRI, Vol.40 (2011), No. 2

Steel: ; low-toughness steel, L-D Steel: )and their

mean values (H-D Steel:
; L-D Steel:
). The
figure also shows the standard value for the Charpy
absorbed energy, i.e., the standard value 27 J has been
represented by a dashed line, the standard value 47 J is
represented by a dotted line and the proposed value 70 J
is represented by a solid line.
In the case of the high-toughness steel, the Charpy
absorbed energy declined significantly in single-pass
welding to an average of 47 J or less. However, it
increased in multi-layer welding to a level equivalent to
that of the base material.
In contrast, in the case of the low-toughness steel,
the Charpy absorbed energy remained almost the same in
single-pass welding at less than 27 J. However, the
Charpy absorbed energy in multi-layer welding
increased significantly to approximately 70 J.
Such recovery and increase in the absorbed energy
has also been reported in some steel weld zones whose
toughness decreased due to cold press bending8) or
aging4).
As described above, it was found that the Charpy
absorbed energy of the weld interface in the case of
high-toughness steel decreased significantly to 47 J or
below in single-pass welding. Although it increased
significantly in multi-layer welding in both the
high-toughness and the low-toughness steels, it was

approximately 70 J even in the low-toughness steel.

3.2 Effect of the order of maximum achieved
temperature
In the weld interface near the middle layer in the
case of multi-layer welding, the achieved temperature at
each pass increases gradually, and reaches 1350 C when
the nearest layer is welded (when the newest weld is
formed) and the achieved temperature then declines in
subsequent passes. In the weld interface of the last
layer, the achieved temperature at each pass increases
gradually as welding progresses, and the maximum
achieved temperature reaches 1350 C when the last
layer is welded. In this way, the order of increase and
decrease of the achieved temperature differs depending
on the position of the weld interface. It is known that
the local toughness in multi-layer welding is affected by
these maximum temperatures achieved at each pass.
Therefore, to study the effect of the order of increase and
decrease of the achieved temperature on the Charpy
absorbed energy, the synthetic HAZ test was also
conducted for the heat cycle of the achieved
temperatures in the order of 8801350880 C
(IntermediateHighIntermediate) and 5008801350 C
(LowIntermediateHigh).
IntermediateHigh
Intermediate and LowIntermediateHigh modeled the
heat cycle of the weld interface of the middle layer and
the last layer, respectively, by rearranging the order of
the achieved temperatures of the heat cycle shown in Fig.
8 (b), which modeled the heat cycle of the weld interface
near the first layer with the achieved temperatures in the
order of 1350880500 C (HighIntermediateLow).
Figure 10 (a) and (b) show each heat cycle.
Figure 11 shows the relationship between the order
of increase and decrease of the achieved temperature at
each pass and the Charpy absorbed energy.
In the high-toughness steel, the Charpy absorbed
energy declined when the order was Intermediate High
Intermediate, which simulated the HAZ of the middle
layer.  However, the Charpy absorbed energy was
approximately 110 J, which is higher than 70 J.  On

1600
1400

Temperature ()

1200
1000
800
600
400
200
0
0

100

200

300
TIME (s)

400

500

600

(a) Intermediate-High-Intermediate
Absorbed energy (J)

1600
1400

Temperature ()

1200
1000
800

250

H-D Steel
L-D Steel

200
150

600

100

400

70J
47J

200

50

27J

0
0

200

400
TIME (s)

600

800

First layer

(b) Low-Intermediate-High

Middle layer

Last layer

(High-Intermediate-Low) (Intermediate-High-Intermediate) (Low-Intermediate-High)

Fig 11 Relationship between the order of achieved

temperatures and the Charpy absorbed energy

Fig 10 Simulated welding heat cycle

(Effect of the order of achieved temperatures)

73

Effect of Heat Cycle in Multi-layer Welding on Charpy Absorbed Energy

nearest layer is welded, significantly influences the

amount of absorbed energy in the weld interface.
Therefore, the authors also conducted the synthetic HAZ
test with the heat cycle in which the achieved
temperature of the first pass (880 C), and that of the
second pass (1350 C) were the same, but the achieved
temperature of the third pass was changed to 1350 C
(IntermediateHighHigh) and 500 C (Intermediate
HighLow). Results of Charpy impact test were
compared to that with the heat cycle in which the
temperature of the third pass was 880 C
(IntermediateHighIntermediate). Figure 12 shows
the heat cycles.
Figure 13 shows the relationship between the
temperature after the maximum temperature was
achieved and the Charpy absorbed energy.
In the case of the high-toughness steel, when the
temperature after the maximum temperature was
achieved was 1350 C, the absorbed energy was slightly
higher than 47 J, which was lower than that of the base
material, as it was in the case of LowIntermediateHigh,
which simulated the weld interface in the last layer.
However, when the temperature after the maximum
temperature was achieved was 500 C, the absorbed
energy was about 110 J, as when the temperature after
the maximum temperature was achieved was 880 C.
On the other hand, in the case of the low-toughness
steel, when the temperature after the maximum
temperature was achieved was 1350 C, the absorbed
energy remained low at approximately 25 J, as in the
case of the LowIntermediateHigh profile, which
simulated the weld interface of the last layer. When the
temperature after the maximum temperature was
achieved was 500 C, the absorbed energy also remained
low at approximately 20 J. Thus, it can be said that
when in multi-layer welding is carried out on the
low-toughness steel used in this study, the Charpy
absorbed energy increases to 70 J or above only when
the achieved temperature is 880 C after the maximum
achieved temperature of the weld interface of 1350 C is
reached.
As described above, in the case of the heat cycles
used in this study, the Charpy absorbed energy of the
weld interface was greatly influenced by the temperature
after the maximum temperature was achieved. In the
case of the high-toughness steel, the Charpy absorbed
energy was recovered when the achieved temperature
was 500880 C, while in the case increased only when
the said temperature was approximately 880 C.

the other hand, when the order was LowIntermediate

High, which simulated the HAZ of the last layer, the
Charpy absorbed energy declined significantly to about
47 J.
In the low-toughness steel, the absorbed energy
exceeded 70 J when the order was IntermediateHigh
Intermediate, which simulated the HAZ of the middle
layer as same as the absorbed energy when the order was
HighIntermediateLow, which simulated the heat cycle
of the HAZ of the first layer. On the other hand, when
the order was LowIntermediateHigh, which simulated
the HAZ of the last layer, the energy remained low at
approximately 20 J.
Thus, it was observed that even in the case of
multi-layer welding and irrespective of the toughness
of steel, the absorbed energy remained low when the
maximum temperature of 1350 C is achieved in the final
pass, as in the case of the weld interface in the last layer.
3.3 Effects of temperature after the maximum
temperature is achieved
On the basis of the previous results, it is considered
that the temperature cycle of the pass after being exposed
to a temperature of 1350 C, which is achieved when the
1600
1400

1,350qC
(Intermediate-High-High)

Temperature ()

1200
1000

880qC

800

(Intermediate-High-Intermediate)

600

500qC
400

(Intermediate-High-Low)

200
0
0

100

200

300

400

500

600

Time (s)

Absorbed energy (J)

Fig. 12 Simulated welding heat cycles

(Effect of the achieved temperature
after the maximum achieved temperature)
200

H-D Steel
L-D Steel
150

100
70J

4. Conclusion

47J

50

27J
0

1350qC

880qC

The following conclusions were drawn in this study.

(1) In the case of single-pass welding, the Charpy
absorbed energy declined significantly in the weld
interface of high-toughness steel. However, in the
case of multi-layer welding, the Charpy absorbed
energy increased more than in the case of

500qC

(High-Intermediate-Low) (Intermediate-High-Intermediate) (Low-Intermediate-High)

Fig. 13 Relationship between the achieved temperature

of the final pass and the Charpy absorbed energy

74

Transactions of JWRI, Vol.40 (2011), No. 2

2)

single-pass welding in some heat cycles, both in

high-toughness steel and low-toughness steel.
(2) On the basis of the heat cycle, it was found that
when the maximum temperature of 1350 C was
reached in the final pass, the Charpy absorbed
energy remained low irrespective of the steel
toughness.

3)

(3) The Charpy absorbed energy of the weld interface

was greatly susceptible to the achieved temperature
after the maximum achieved temperature, and the
range of the temperature after the maximum
temperature was achieved in which the Charpy
absorbed energy was recovered was found to be
different depending on the steel toughness.

4)

5)

6)

Acknowledgments
This study was financially supported in part by the
research grant from the Japan Iron and Steel Federation.
The authors would like to thank Mr. Nakatsuji and Dr.
Tamagawa for their help during the experiments and the
simulation.

7)

References
1)

H. KUWAMURA, H. AKIYAMA, S. YAMADA and

J.-C.CHIU: Experiment on the mechanical properties and
their improvement of cold press-bent steel plates, Journal
of Struct. Constr. Eng., (1993) No.444, pp.125-133. (in
Japanese)

8)

75

Y. SAKINO, S. TAKAHASHI and Y.-C. KIM: Effects of

Strain Rate on Tensile Strength of Steel Specimens of
HAZs with Stress Concentrations, Quarterly Journal of the
Japan Welding Society, Vol.28 (2010) No.3, pp.328-337.
(in Japanese)
Y. SAKINO, H. KAMURA and Y.-C. KIM: Charpy
Absorbed Energy at CGHAZ of Rolled Steels for Building
Structures Based on Simulated Heat-input, Journal of
Constructional Steel, Vol. 13 (2005), pp.157-164. (in
Japanese)
Y. SAKINO, H. KAMURA and Y.-C. KIM: Effect of
Pre-Strain and Aging on Charpy Absorbed Energy in
Welding Heat Affected Zone, Journal of Structural
Engineering, Vol.52B (2006), pp.327-334. (in Japanese)
Y. SAKINO, H. KAMURA and Y.-C.KIM: Effect of
Welded Condition on Charpy Absorbed Energy of Heat
Affected Zones in Low Toughness Steel, Steel
Construction Engineering, Vol.17 (2010) No.67, pp.43-52.
(in Japanese)
M. MOCHIZUKI and M. TOYODA: Effects of Welding
Multiple Heat Cycles on Joint Performance in
Beam-to-Column Welded Connection of Steel Framed
Structures Studies on Relation between Joint
Performance and Welding Heat Cycles in Welded Joints
(Report 1), Quarterly Journal of the Japan Welding
Society, Vol.21 (2003) No.1, pp.46-53. (in Japanese)
Y.-C.KIM, J.-Y LEE and K. INOSE: High Accurate
Prediction of Welding Distortion Generated by Fillet
Welding, Quarterly Journal of the Japan Welding Society,
Vol.23 (2005) No.3, pp.431-435. (in Japanese)
H. KUWAMURA and Y. MATSUMOTO: Fracture
Resistance of 800-MPa Steel Subjected to Heat Cycles,
Journal of Struct.Constr.Engng. , No.484 (1996),
pp.101-109. (in Japanese)